feat(models): cross-plane promotion decision + RunReceipt (fail-closed, replayable) - #274
Merged
Conversation
…d, replayable) decide_promotion() is the one promotion contract both planes call: a challenger promotes only if it beats the champion AND passes the eval gate; regulated targets need examiner sign-off; nothing self-promotes (eval_passed/signed_off default False). Emits a RunReceipt whose hash is deterministic over the decision inputs — a promotion is examiner-replayable. 8 tests + a tools-tests CI workflow (gates promotion + surface-feed tests, previously ungated).
8 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The governance capstone — closes the champion/challenger loop with a deterministic, examiner-replayable promotion.
tools/promote_model.py—decide_promotion(), the one promotion contract both planes (cloud lattice + local agent-machine) call. A challenger promotes only if it beats the champion and passes the eval gate; a regulated target additionally needs examiner sign-off; nothing self-promotes (eval_passed/signed_offdefault False). Verdict ∈ promote|hold|shadow|blocked, plus a PSI-drift flag.RunReceiptwhose hash is deterministic over the decision inputs — re-run the inputs, get the same receipt. Examiner-replayable.tools-testsCI workflow that gates the promotion + surface-feed tests (previously ungated). Cross-linked from the governed-model-plane doc.Verified: 17 tools tests pass together (promotion + feeds).